Assessing the Merit of Soil Moisture as both a Metric and Predictor of Drought for Colorado and the Upper Colorado River Basin Using Land Data Assimilation Models.

Abstract:

Root zone soil moisture (RZSM) is the water in the soil that is within the reach of surface vegetation. When RZSM becomes sufficiently low plants are no longer able to overcome the suction force holding water in the soil. This skews the partitioning of latent and sensible heating in favor of sensible heating, thus warming the surface, and potentially invoking a positive feedback.

Findings from Koster et al published in 2004 indicate that not only is there potential for the improvement of seasonal forecasts through tracking RZSM, but also that RZSM feedbacks are strongest in what can be thought of as wet-dry transitional zones. These are zones where surface evaporation rates average high enough to be expected to have an important influence on precipitation, and where available soil moisture is still an important constraint on how much surface evaporation takes place. In the Upper Colorado River Basin and eastern Colorado climate varies rapidly with space due to differences in elevation, and these transitional zones do exist within the domain.

This paper focuses on how NASA Land Data Assimilation Modeled RZSM is used to help track drought in the Upper Colorado River Basin and eastern Colorado, and addresses the additive predictive skill RZSM may have on multi-weekly and seasonal timescales across the Upper Colorado River Basin and eastern Colorado during the growing season. Daily modeled soil moisture from three land data assimilation models was correlated with North American Regional Reanalysis temperature data, and precipitation data from rain gauges interpolated using PRISM climatology in order to help answer important questions about the predictive skill of soil moisture, and its value in the drought early warning process.

Questions addressed here will be as follows: In what climate regimes within the domain does RZSM have the most predictive power over temperature and precipitation? Do certain predominant soil and vegetation types preferentially strengthen RZSM-atmosphere interaction? How does the behavior of RZSM during the onset of severe droughts within the domain over the last 20 years compare to normal behavior? How much do these answers vary between land surface models given the same forcing inputs?